If the format requested was not available, we called the method '.get' in None.
Bugfix for bug #4123 & #4153
Also, replace deleted test case
Also, modernize a bit.
fix swrmediathek for new formats.
[generic] indicate when a direct video has been extracted
The generic extraction no longer works.
Also remove width and height as they are not of the actual video.
[channel9] Fix extraction
The timestamp fluctuates with DST.
'<a href="foo" ><img src="bar" / ></a>' wouldn't be parsed right (the problem is '/ >', '/>' worked fine). We need to change it in python 2.6 (for example the description of youtube videos wouldn't be extracted).
It would print the debug headers for each test. And nose uses a StringIO object for stdout, which in python 2.x doesn't have the 'encoding' attribute.